The volume of tweets and Google Search Volume Index (SVI) were found to be leading price indicators for bitcoin and Ethereum, according to a research paper published by the Southern Methodist University.
The importance of sentiment
In the , researchers gathered data on Twitter mentioning bitcoin and Ethereum; the same was done using Google trends. Building on the ideas of previous research, the hypothesis was that the number of tweets and their sentiment (positive and negative) can influence prices. In the study, it was uncovered that the number of tweets and Google searches changes first before prices do.
The role of in technical or market analysis is to uncover people’s attitudes towards an entire market or individual index (in this case bitcoin and Ethereum). The theory of sentiment analysis is a branch of technical analysis that states that price discounts everything, and that price trends are ultimately a reflection of crowd psychology.
Therefore, in theory, if you could measure how positive or negative the people’s shared views are towards a particular stock or cryptocurrency, you could estimate its price trajectory.
Although in this particular study, tweet volume, and not sentiment, was found to be a leading factor in the price of cryptocurrencies. The lack of sentiment being the leading factor was theorized due to the amount of “noise” there is on Twitter about the currencies compared to actual conversation.
For instance, the researchers found that there 21 million bots on Twitter posting mostly factual information about prices, advertisements, spam etc. Not humans having real discussions about how they feel about either bitcoin or Ethereum.
The other issue that researchers found with Twitter was that sentiment was mostly positive in nature — even when the prices of bitcoin and Ethereum were falling.
People who tweet about cryptocurrencies even when their prices drop have an interest in them beyond investment opportunity making the tweets biased towards positive.
Despite their findings, the researchers did not completely rule out sentiment analysis using different modeling techniques.
Methodology
In the study, researchers used to open source (Valence Aware Dictionary and Sentiment Reasoner) for analyzing tweet data. Tweet data was taken dating back to 2014 using the site bitinfocharts.com. Google trends data (SVI) was taken as far back as 2004 scaled in the terms proportion to all searches on all topics for the terms bitcoin and Ethereum.
Results
For the Google trends data, the report found that the price was highly correlated with searches for the keyword bitcoin and Ethereum, and that these search spikes occurred before the actual increase in prices were observed.
Another strong correlation between Twitter and bitcoin’s price was found, except this time with more compelling results.
Finally, using machine learning, the results from the Google trends and tweet data was also put into a linear model to verify the positive correlations. The data was split between a training model and testing in an 80% and 20% split.
Social Media Helps Monitor Investor ‘Chatter’
The VADER data could provide some valuable data for investors in gauging market sentiment.
Previously, Bitcoinist has covered the importance of social media chatter on Twitter before with tools such as the .’ But this is the first time that Twitter and SVI data has shown to lead, and not follow, the prices of the most popular cryptocurrencies.
Can social media data provide valuable trading insight? Share your thoughts below!
Images via Shutterstock
The Rundown
Published at Mon, 06 May 2019 02:00:29 +0000
